[miso-users] newbie question about "“(0,0,0): N" where the N is bigger than any other category for most events.

[Hazard, E. Starr]

Folks,

I have a question about MISO output and you folks are experts.

I have a dataset analyzed and now I am vetting the respective output.

For a candidate gene with three transcripts, I  see from the *bf file:


Sample1_counts                                                                                                                                                 S1_assigned_counts                                                                                            Sample1_counts                                                                                                                 S2_assigned_counts
(0,0,0):3117,(0,0,1):19,(0,1,0):543,(1,0,0):2,(1,0,1):303,(1,1,1):135

0:261,1:639,2:102

(0,0,0):3049,(0,0,1):21,(0,1,0):640,(1,0,0):2,(1,0,1):241,(1,1,1):127

0:195,1:734,2:102


The assigned_counts per transcript seem to make sense

However, for the assigned counts, the largest numbers are seen for the “(0,0,0): 3117” and “(0,0,0): 3049” cases.

From the on-line manual

“The field (0,0):N, if outputted, indicates the number of reads N that mapped to the region but were thrown out (e.g. because they are not consistent with the isoforms in the annotation, violate overhang constraints, or other related reasons.)”


1) Have you seen this before?
2) How can I evaluate whether or not that result is correct? That is, am I really throwing out 3K reads per the manual? Remember this is for a single event/Gene. If so I am losing a lot of data.
3) How do you handle replicates? Miso works on pairs.


Cheers,

Starr





-------------------------------------------------------------------------
This message was secured via TLS by MUSC.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/miso-users/attachments/20170922/3e6a0bff/attachment.html